Scribd Logo
Upload

Welcome to Scribd!

  • 36 views

    Assignment 1

    Uploaded by

    Vivek Chanpa
    This document provides instructions for Assignment #1 for course MM531. It includes 3 problems to work on: 1) writing code to rearrange and transform a user-input matrix, 2) performing linear regression on dataset "MM531data.xlsx" by writing code from scratch to normalize data, define functions, and use gradient descent, and 3) using logistic regression to classify crystal structures in "HEA_phases.xlsx" by writing custom code without libraries to clean data, define numeric labels, state functions, and use gradient descent. Students are asked to submit their code and output by the due date.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd

    Assignment 1

    Uploaded by

    Vivek Chanpa
    36 views2 pages
    This document provides instructions for Assignment #1 for course MM531. It includes 3 problems to work on: 1) writing code to rearrange and transform a user-input matrix, 2) performing linear regression on dataset "MM531data.xlsx" by writing code from scratch to normalize data, define functions, and use gradient descent, and 3) using logistic regression to classify crystal structures in "HEA_phases.xlsx" by writing custom code without libraries to clean data, define numeric labels, state functions, and use gradient descent. Students are asked to submit their code and output by the due date.

    Original Title

    ASSIGNMENT 1

    Copyright

    © © All Rights Reserved

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document provides instructions for Assignment #1 for course MM531. It includes 3 problems to work on: 1) writing code to rearrange and transform a user-input matrix, 2) performing linear regression on dataset "MM531data.xlsx" by writing code from scratch to normalize data, define functions, and use gradient descent, and 3) using logistic regression to classify crystal structures in "HEA_phases.xlsx" by writing custom code without libraries to clean data, define numeric labels, state functions, and use gradient descent. Students are asked to submit their code and output by the due date.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    36 views2 pages

    Assignment 1

    Uploaded by

    Vivek Chanpa
    This document provides instructions for Assignment #1 for course MM531. It includes 3 problems to work on: 1) writing code to rearrange and transform a user-input matrix, 2) performing linear regression on dataset "MM531data.xlsx" by writing code from scratch to normalize data, define functions, and use gradient descent, and 3) using logistic regression to classify crystal structures in "HEA_phases.xlsx" by writing custom code without libraries to clean data, define numeric labels, state functions, and use gradient descent. Students are asked to submit their code and output by the due date.

    Copyright:

    © All Rights Reserved
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 2

    MM531 Assignment #1

    Note: If you are a BTech student enrolled in MM431, then you must NOT
    submit your answers for this problem set. This is only for students enrolled
    in MM531.

    Please submit the screenshots of the output (as a single pdf file), the revised
    databases and the python files (.py or .ipynb) via Google Classroom by
    Thursday midnight.

    Problem 1:
    Write a code for entering a m x n matrix (user to enter the values of m and n) from
    the console. First re-arrange the matrix such that all the row elements in a given row
    are in increasing order. Then reverse all the elements of each individual column (i.e.,
    elements at top goes to bottom and so on). Print out the transformed matrix.
    (a) Run the code with a set of input values of your choice.
    (b) Carry out the transformations manually and compare with the computer
    output. Show your work.

    Problem 2:
    Read the dataset “MM531data.xlsx”. Normalize the y values but not the x values. You
    have to perform a linear regression by writing the code from scratch without using
    scikit learn.
    (a) Process your data by first visualizing and then normalizing only the y
    values. Save your revised database as MM531data_rev.xlsx.
    (b) State your hypothesis function and loss function.
    (c) Write the code for performing linear regression with the code being
    optimizing using the Gradient Descent Method. The code should be written
    such that the code will automatically update the learning rate instead of you
    having to do this manually. Hint: This is similar to tutorial #3 except that
    instead of using a single loop, you use a nested loop with the outer loop
    performing the learning rate optimization and the inner loop performing the
    gradient descent.
    (d) Start with an initial estimate for intercept as 0.1 and slope as 0.01, as well
    as a learning rate of 0.01 for both slope and intercept. Show after running
    the code that your optimization converges. How many iterations were needed
    for optimizing the learning rate (i.e., how many times did you have to iterate
    in the outer loop before convergence)? What are the optimized values of
    slope and intercept?
    Problem 3:
    Read the dataset “HEA_phases.xlsx” attached with this assignment and process it
    (drop the zeros and nulls). You will need to write a code from scratch (i.e., without
    using any of the ML libraries) in order to train a Logistic Regression model with all of
    the given data (i.e., use all of it for training). Perform the following steps.
    (a) Explore your dataset and get information about your dataset.
    (b) Clean up the dataset – get rid of nulls, zeros and negatives. Save the new
    dataset as HEA_phases_rev.xlsx
    (c) Write a simple script such that the crystal structures can be treated in
    numeric format. Overwrite HEA_phases_rev.xlsx created in part (c).
    Here, you basically need to assign a number value (1 or 2) according to
    the two crystal structures in the system.
    (d) Write down the hypothesis function (i.e., logistic regression) and the loss
    function that you will be minimizing.
    (e) Write a script for classification of crystal structures as FCC or BCC
    using Logistic Regression from scratch (i.e., without using ML libraries
    such as scikit learn etc.). Use the Gradient Descent method for
    optimizing the Logistic Regression. Write down the expression for the
    optimized hypothesis function.

    You might also like